Soulful, Pure, Benevolent. #SPB #SPBalasubrahmanyam

This prescient warning from Dr Manmohan Singh in 1999 should serve to caution the bigots in our ruling circles — and also to inspire the rest of us to work to prevent the worst!

JHU CLSP researchers have co-authored (at least) 19 papers at #ICASSP2021 between them. Together, these represent collaboration with authors from 18 different affiliations. Here's a thread with dates📅 and links to videos📽️ and papers📎. (All times are in EDT.)

Our lab will be presenting 20 papers at #INTERSPEECH2021! This is also our inaugural tweet 👋 Follow for updates future updates on our research activities in **end-to-end speech** recognition, enhancement, translation, etc. + open-source developments on ESPnet.

📣We are hiring one more researcher: please apply to work with us on the cutting edge of audio, speech, and language processing research! ⬇️More details and how to apply. 8/end merl.com/employment/emp…

Thanks a lot Jonathan Le Roux. I am delighted to join the amazing speech & audio team Mitsubishi Electric Research Laboratories 🎉

We're happy to announce that, now that #WASPAA2021 will be fully virtual, we have discounted all students registrations to $20! Registration is also *free* for low-income countries. Register at edas.info/registerPerson… (you'll need to sign in/up to EDAS). RT welcome.

Our lab is looking for a post-doc in ASR-related research. Please find more details in cmu.wd5.myworkdayjobs.com/en-US/CMU/job/… We are looking forward to working with you🤝🤝🤝

I just made a Google colab based tutorial for espnet, and used it at my course in CMU. colab.research.google.com/github/espnet/…

We have a new paper out introducing the "cocktail fork problem", where we bite into a small chunk of the cocktail party problem by separating real-world soundtracks into 3 stems: music, speech, and sound effects. Example video and more details below. 👇🧵

Join us on Dec. 9 for overviews of MERL's research and to chat with researchers on Gather.Town. I'll have the pleasure to introduce recent work by the Speech & Audio team.

Want to learn to separate sounds based on non-mutually exclusive conditions? Or maybe you prefer to generalize to unseen in-domain discriminative concepts?🧐 Check our new paper with Jonathan Le Roux and Aswin Subramanian from Mitsubishi Electric Research Laboratories! Code & data coming soon: github.com/etzinis/hetero…

Let's register yourself or your colleagues to make a richer reviewer pool for SLT! slt2022.org/call-for-revie…

King Leo ends the GOAT debate once and for all 🤩

New paper out w/ Christoph Boeddeker, Aswin Subramanian, Gordon Wichern, Reinhold Haeb-Umbach: "TS-SEP: Joint Diarization and Separation Conditioned on Estimated Speaker Embeddings", initiated during Christoph's internship Mitsubishi Electric Research Laboratories last year. arxiv.org/abs/2303.03849

Congrats Darius on receiving an #ICASSP2023 IEEE ICASSP Best Student Paper Award 🥳🎉

"Hyperbolic audio source separation" by Darius Petermann, Gordon Wichern, Aswin Shanmugam Subramanian, Jonathan LeRoux ieeexplore.ieee.org/document/10094…

India's great contradiction is that the shortest belt accounting for 50% of India's economic activity has no overlap with the shortest belt accounting for 50% of population